# OpticParse & PhishVision -- AI Agent Web Mesh & x402 Paywall

This file is provided for LLMs (ChatGPT, Claude, Devin, Perplexity, ElizaOS)
exposing standardized, machine-readable extraction and threat scanning actions.

## 1. Vision Web Scraping (Cloudflare Bypass & Multimodal Vision)
- Endpoint: POST https://opticparse-edge.parastejpal987.workers.dev/api/edge/scrape
- Autonomous Starter Trial: POST https://opticparse-api.onrender.com/api/auth/bot-key (payload: {"bot_name": "MyAgent", "framework": "ElizaOS", "owner_contact": "agent@domain.com"} -> 100 free requests)
- Autonomous Paywall: HTTP 402 via X-Payment-TxHash ($0.01/call settled via $0.05 USDC micro-pack [5 credits] on Base, Polygon, or Arbitrum to 0xd458E709e7d54fd3659EF66624A621Cde74EDD27)
- Machine Discovery Standard: https://opticparse.com/.well-known/agent.json (Google A2A v1.0)
- Request Schema:
{
  "url": "https://example.com",
  "query": "Extract structured fields in JSON"
}

## 2. PhishVision 0-Day URL Threat Detector
- Endpoint: POST https://opticparse-api.onrender.com/api/phish-detect
- Schema:
{ "url": "https://suspicious-domain.com" }

## 3. Clean Markdown & RAG Parser (Token-Optimized for LLMs)
- Endpoint: POST https://opticparse-api.onrender.com/api/extract/markdown
- Request Schema:
{
  "url": "https://example.com/article",
  "include_images": false,
  "include_links": true
}
- Response: Returns pristine, noise-stripped Markdown text with token count and title.

## 4. Interactive Browser-Use Agent
- Endpoint: POST https://opticparse-api.onrender.com/api/agent/interact
- Request Schema:
{
  "target_url": "https://example.com",
  "actions": [
    {"action": "click", "selector": "#button"},
    {"action": "type", "selector": "#input", "text": "value"}
  ],
  "final_query": "Extract data"
}

## 5. Real-Time Zero-Day Threat Stream
- WebSocket: wss://opticparse-api.onrender.com/ws/threat-stream
- Real-time JSON broadcast of newly discovered phishing domains and malicious kits.

## 6. Python SDK (Instant Integration)
```python
from opticparse import OpticParseClient
client = OpticParseClient(api_key="dev_key_for_agents")
result = client.scrape("target_url", "Extract pricing data")
```